Page 1 of 1

Designate missing values

Posted: Tue Jan 30, 2018 11:49 am
by hejhej
First of all, thank you for a great piece of software!

I have a suggestion: make it possible to assign some values as missing in the "Data" view. This would be very useful, espacially when importing large data files from other programs that have discrete values as missing.

Re: Designate missing values

Posted: Wed Jan 31, 2018 12:28 am
by jonathon
hi,

can you describe what you mean by "that have discrete values as missing"?

with thanks

Re: Designate missing values

Posted: Wed Jan 31, 2018 5:49 pm
by superyms
I think he is proposing to populate missing data with NA, like we could do when importing csv files in R using na.strings=NA

That would be nice to be done in Jamovi.

Re: Designate missing values

Posted: Wed Jan 31, 2018 11:42 pm
by jonathon
hi yes,

a basic version of this is in the latest release (0.8.1.17) (currently available for macOS and linux, and the windows version should arrive in the next couple of hours).

you can set the 'default missings' in the app menu. you can specify NA, or -999999, and this will be treated as a missing value. however, this only applies at import time -- once the data is imported this has no effect.

we'll add the ability to change the missing value value on an *existing* data set in the future.

cheers

jonathon

Re: Designate missing values

Posted: Thu Feb 01, 2018 6:57 am
by hejhej
Thanks, the kind of feature I was looking for! I have an SPSS file where missing is set as 9, 99 or in some cases 999 and that is giving me a bit of a headache when trying to do analysis in Jamovi.

Re: Designate missing values

Posted: Thu Feb 01, 2018 7:01 am
by jonathon
hmmm,

at present, there's only provision for a single missing value value :/

but you could use the new `IFELSE` compute function:

IFELSE(colName == 9, NA, colName)

jonathon

Re: Designate missing values

Posted: Thu Feb 01, 2018 8:45 am
by jonathon
In the next release you'll be able to use FILTER, which is almost the same:

FILTER(colName, colName != 9)

it's just a bit more "descriptive"

jonathon